7

LYME ARTHRITIS AN EPIDEMIC OF OLIGOARTICULAR ARTHRITIS IN CHILDREN AND ADULTS IN THREE CONNECTICUT COMMUNITIES ALLEN C. STEERE, STEPHEN E. MALAWISTA, DAVID R. SNYDMAN, ROBERT E. SHOPE, WARREN A. ANDIMAN, MARTIN R. ROSS, and FRANCIS M. STEELE An epidemic form of arthritis has been occurring in eastern Connecticut at least since 1972, with the peak incidence of new cases in the summer and early fall. Its identification has been possible because of tight geographic clustering in some areas, and because of a characteristic preceding skin lesion in some patients. The authors studied 51 residents of three contiguous Connecticut communities-39 children and 12 adults-who developed an From the Departments of Internal Medicine. Pediatrics, and Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut; the Field Services Division, Bureau of Epidemiology, Center for Disease Control, USPHS, Atlanta, Georgia; and the Preventable Disease Division and Laboratory Division, State Health Department, Hartford, Connecticut. Presented in part at the 40th Annual Meeting of the American Rheumatism Association, Chicago, Illinois, June 30, 1976 ( I ) . Supported in part by USPHS grants AM-10493. AM-5639, AI-10984, BRSG-RR-05443, and RR-00125, by the Connecticut Chapter and National Office o f The Arthritis Foundation, and by the Kroc Foundation. Allen C. Steere, M.D.: Postdoctoral Fellow in Rheumatology, Department of Internal Medicine, Yale University: Stephen E. Malawista, M.D.: Head. Rheumatology Section, Departmentbf Internal Medicine, Yale University: David R. Snydman, M.D.: Acting Director, Preventable Disease Division, State Department of Health, and Bureau of Epidemiology, Center for Disease Control; Robert E. Shope, M.D.: Department of Epidemiology and Public Health, Yale University; Warren A. Andiman, M.D.: Department of Pediatrics, Yale University; Martin R. Ross, Ph.D.: Laboratory Division, State Department of Health; Frances M. Steele, Ph.D.: Laboratory Division, State Department of Health. Address reprint requests to Allen C . Steere. M.D., Section of Rheumatology, Department of Internal Medicine. Yale University School of Medicine, New Haven, Connecticut 06510. Submitted for publication September 16, 1976 accepted September 18, 1976. Arthritis and Rheumatism, Vol. 20, No. I (January-February 1977)

illness characterized by recurrent attacks of asymmetric swelling and pain in a few large joints, especially the knee. Attacks were usually short (median: 1 week) with much longer intervening periods of complete remission (median : 2.5 months), but some attacks lasted for months. To date the typical patient has had three recurrences, but 16 patients have had none. A median of 4 weeks (range: 1-24) before the onset of arthritis, 13patients (25%)noted an erythematouspapule that developed into an expanding, red, annular lesion, as much as 50 cm in diameter. Only 2 of 159 family members of patients had such a lesion and did not develop arthritis (P < 0.000001). The overall prevalence of the arthritis was 4.3 cases per 1,000 residents, but the prevalence among children living on four roads was 1 in 10. Six families had more than 1 affected member. Nine of 20 symptomatic patients had low serum C3 levels, compared to none of 31 asymptomatic patients ( P < 0.005); no patient had iridocyclitis or a positive test for antinuclear antibodies. Neither cultures of synovium and synovial fluid nor serologic tests were positive for agents known to cause arthritis. “Lyme arthritis” is thought to be a previously unrecognized clinical entity, the epidemiology of which suggests transmission by an arthropod vector.

I n November 1975 a mother from Old Lyme, Connecticut, informed the State Health Department that 12 children from that small community of 5,000, 4 of whom lived ciose together on the same road, had a disease diagnosed as juvenile rheumatoid arthritis (JRA). During the same month another mother from

STEERE ET AL

8

the same community reported at the Yale Rheumatology Clinic and to the Health Department that she, her husband, 2 of their children, a n d several neighbors all had arthritis. Again most of the children were thought t o have JRA. Subsequent studies of children and adults in that geographic region suggest that “Lyme arthritis” is a previously unrecognized clinical entity.

MATERIALS AND METHODS A system of surveillance was organized in the three contiguous communities of Old Lyme, Lyme, and East Haddam (total population: 12,000) through contacts with healthcare personnel-area physicians, school nurses, and local health officers. In addition, several patients identified other affected individuals. In these small communities, such methods are likely to have identified at least those patients severely enough affected to have sought medical assistance, and through the patient “grapevine,” to have emphasized the presence of family, geographic, or social clustering. in an effort to exclude children with diverse forms of arthritis who may have clustered by chance alone, affected residents were invited to participate in a study at the Yale University School of Medicine, in which an attempt would be made to determine the cause of their arthritis.* From December 1975 through May 1976, the same physician evaluated all patients by means of an extensive history, a physical examination, blood tests, and if possible a synovial biopsy. In addition, the authors contacted the patients’ physicians (many patients had several-general practitioner, internist, pediatrician, rheumatologist, orthopedic surgeon, or opthalmologist), to study their office records and available hospital records. The study included only patients with both joint pain and swelling, documented by a physician and diagnosed as J R A or arthritis of unknown etiology. Of the 51 resulting patients, 2 0 had active disease at the time of the author’s examination. Laboratory tests performed on all patients in the study (except as noted) included peripheral blood and synovial fluid cell and differential counts, erythrocyte sedimentation rates (Westergren), and synovial fluid protein and glucose determinations. Both blood and joint fluid specimens were tested for rheumatoid factor by latex fixation, for antinuclear antibodies by immunofluorescence, and for the third component of complement (C3) by radial immunodiffusion. In addition, serum antibodies against various agents were sought in all patients as follows: adenoviruses, cytomegalovirus, herpes simplex, influenza, and mumps viruses, by complement fixation; rickettsiae by complement fixation and immunofluorescence (the latter, in 10 sera); M pneumoniae, by complement fixation and cold agglutination; groups A, B, and other arboviruses, by complement fixation and hemagglutination inhibition (2,3); leptospira by microscopic agglutination (in 11 sera) (4); rubella by hemagglutination inhibition (also rubella-specific IgM antibodies were sought in 16 sera in which IgG had been adsorbed to Staphylococcus aureus (5,6); and coxsackie viruses

* Protocol No.

1125, approved by the Human Investigations Committee, Yale University School of Medicine.

BI-5 and A-9, by neutralization tests done in rhesus monkey kidney tissue cultures (7). Hepatitis B surface antigen was sought both by radioimmunoassay and by reversed passive hemagglutination; antibody (in 16 sera) was sought by passive hemagglutination (Abbott Laboratories, North Chicago, IL). Twenty patients (I6 children and 4 adults) were tested for the histocompatibility antigen HL-A B27 (8). Synovial fluid specimens were inoculated in thioglycolate broth and on blood, deoxycholate, chocolate, and pleuropneumonia-like organism (PPLO) agar (Difco Laboratories, Detroit, M I ) supplemented with dextrose, yeast extract, and horse serum. The PPLO plates were incubated in a C02-rich environment at 37°C for 3 weeks. Specimens of synovium and joint fluid were each placed in tubes containing a monolayer of human placental fibroblasts and Eagle’s basal medium supplemented with 10% fetal calf serum. After transport to the laboratory, each specimen was inoculated into four tissue culture systems: human placental fibroblast, Vero, Hep-2, and rhesus monkey kidney. Each specimen was examined for cytopathic effect twice weekly for 4 weeks. Throat and rectal swab specimens were taken from the 7 patients who had the onset of arthritis during the study, and these specimens were cultured for viruses in the same manner.

RESULTS Clinical Characteristics Fifty-one residents (39 children and 12 adults) were identified who had an apparently similar type of arthritis. It usually began with the sudden onset of swelling, and often pain, in a knee, another large joint, or the temporomandibular joint (Table 1). Five patients also described pain in unspecified joints of the hands or feet. The onset was monoarticular in 35 patients (69%), oligoarticular in 15 (29%), and polyarticular in only 1 (2%). T h e first attack lasted a median of 1 week, but sometimes persisted for as long as 6 months (Table 2). To date, 35 patients (69%) have had recurrent attacks, and t h e typical patient has h a d three recurrences (range: 2-10). Large joints, particularly the knee, were again most commonly affected, and a median of three joints, usually one a t a time, were affected in all attacks (Table 1). Recurrences were usually short (median: 1 week) and separated by much longer periods of complete remission (median: 2.5 months) (Table 2). However, even in an individual patient, both the duration of recurrent attacks (up to 4 months) a n d t h e interval between them ( u p t o 23 months) were highly variable and therefore unpredictable. During remissions some patients remembered short periods of joint pain, sometimes lasting only hours, without swelling (and therefore not included as attacks). Children and adults did not differ significantly in joint symptoms. Although nearly half the patients reported only

LYME ARTHRITIS

9

Table 1. Joints Affected Number of Patients First Attack Children (39)

Adults (12)

35 3 3 2 2 3 2

8 2 2 3 2 I

Recurrent Attacks Total (51)

Children (25)

Adults

22 5 7 2 4 4

8 3

(10)

Total (35) ~

Knee Ankle Wrist Temporomandibular Shoulder Hip Elbow

* Percentage

43 5 5 5 4 4 3

I

(84)* (10) (10) (10) (8)

(8) (6)

10

1

3 3 0 6

30 (86)* 8 (23) 8 (23) 5 (14) 7 (20) 4 (11) 16 (46)

o f patients with either or both joints affected

articular symptoms, 28 had concomitant fever (100103"F), 14 malaise and fatigue, 8 headache, 8 myalgia, and 3 a maculopapular rash. Several adults described episodic symptoms not necessarily associated with documented arthritis, such as severe headache, periorbital edema, malar or photosensitive maculopapular rash, or swelling of the hands or feet. In addition, 7 adults noted profound fatigue and hyperesthesias, sometimes persisting for months after the arthritis had gone. During the study period from December 1975 through May 1976,20 of the 51 patients were symptomatic, usually with recurrent attacks, at the time of examination. In these patients the typical physical finding was marked swelling and sometimes warmth of a single joint, with mild to moderate pain on motion. N o one was found t o have evidence of permanent joint damage. None of the 22 children who had had ophthalmologic examinations was found to have iridocyclitis.

Preceding Symptoms A median of 4 weeks (range: 1-24) before the onset of arthritis, 13 patients (25%; 8 adults and 5 children) described an erythematous papule that developed into an expanding, red annular lesion, usually with partial central clearing. The lesion, often on an extremity, became 20-50 cm in diameter in some patients. Three patients had multiple concurrent lesions, but none had recurrent lesions. Associated symptoms included burning of the lesion in 10 patients, fever in 7, severe headache in 5 , stiff neck in 3, nausea and vomiting in 2, and Bell's palsy in 1 . The skin lesion lasted a median of 1.5 weeks (range: 0.5-4). Although physicians and patients thought it to be an insect bite, only 1 patient remembered having been bitten at the site of the lesion, in that instance by a tick. Adults had the skin lesion significantly more often than children (P < 0.01). Only 2 of

Table 2. Clinical Characteristics

First attack Type o f onset Monoarticular Oligoarticular (1-4 joints) Polyarticular (5 or morejoints) Duration of first attack (weeks) Recurrent attacks Number of patients Number of recurrent attacks joints affected in all attacks Duration of recurrent attacks (weeks) complete remissions (months)

*

Median (range).

Children (39)

Adults (12)

Total

28 (72%) 10 (26%) I ( 2%) 1 (.3-12)*

7 (58%) 5 (42%) 0 I .5 (1-24)

35 (69%) 15 (29%) I ( 2%) I (.3-24)

25 (64%)

10 (83%)

35 (69%)

3 (1-6) 3 (1-7)

2.5 (1-10) 4 (1-8)

3 (1-10) 3 (1-8)

I (.l5-12) 2 (.25-23)

I (1-16) 3 (.5-ll)

(51)

1 (.15-16) 2.5 (.25-23)

STEERE ET AL

10

the 159 family members of patients noted such a lesion and did not develop arthritis (P< 0.000001). No patient described a sore throat or diarrhea before the onset of arthritis.

Epidemiology The three communities studied are all small (population 5,400 or less) and relatively sparsely settled (each area 29 square miles or more). Old Lyme borders Long

Island Sound, all three communities border the Connecticut River, and the interiors are laced with streams and lakes (Figure 1). Except for the town centers of Old Lyme and East Haddam, most residents live on large wooded lots or on farms. Each household has its own well water. The residents of Old Lyme and Lyme share a school system; those of East Haddam have a separate one. The onset of arthritis in the 51 patients occurred from July 1972 through May 1976 (the cut-off time for

Long

I s l a n d Sound

Fig 1. The three communities studied are shown. Two state forests and major lakes are indicated. bur the enrire region is laced wirh smaller lakes and streams. Geographic clustering occurred in rhe more sparsely settled. heavily wooded areas, and not in the town cenrers or along Long Island Sound.

LYME ARTHRITIS

11

this study), except for 1 patient whose arthritis began in March 1967.* The 44 patients whose arthritis began from 1972 through 1975 are represented in Figure 2. Twenty-eight patients (55%) had their onset from June through September. All patients with the expanding skin lesions developed them during these months. The onset of arthritis in 6 additional patients was in the first 5 months of 1976 (1 each in January, March, April, and May; 2 in February). At the onset the patients had a median age of 1 1 (ra.nge: 2-45), and the sex ratio was 1.3: 1 in favor of males (Figure 3). The overall period prevalence rate was 4.3 per 1,000 residents (Table 3). However marked geographic clustering was observed. Six families had more than 1 affected member. Half the affected residents in Old Lyme lived on two adjoining country roads, as did half of those affected in East Haddam. One in 10 children living on those four roads had the illness (Table 4). In contrast, no affected resident was identified who lived in the town centers or on Long Island Sound. Residents who lived close together usually did not have the onset of illness at the same time. In all but 1 of the families with more than 1 affected member, those affected had the onset of symptoms in different years. Similarly, patients living on the same roads often had

1972

'1 '1

0 Old

HLyme 0 East Haddom

I

1973

'

m

0

5

1974 I

JAN

SEPT

JULY

MAY

MAR

NOV

MONTH OF ONSET

Fig 2. The months of onset of arthritis and towns of origin are shown for the complete years 1972 through 1975 (44 of the J I patients). Twentyeight patients (55%) had their onset from June through September. and all I 3 patients (25%) with preceding, expanding skin lesions (see text) had them during that 4-month period. Only 2 of the IS9 nonarthritic family members of patients had similar skin lesions ( P < 0.000001 I.

the onset of the illness in different years, as did friends and classmates who might have had contact with one another. Fifteen patients did not know others with the illness. No common exposure, such as an immunization,

* This patient seems to belong to the group because she had symptoms similar to those of her husband and 2 of their 4 children, who currently have arthritis. She is mentioned to indicate the possibility that the disease has been endemic in the area for some time.

NO. OF CASES

10

c

0

n

H

Male

Female Expanding skin lesion

n

Idl I2

56

910

13-

17-

14

18

Lyme

2122

I.1 25-

29-

26

30

3334

3738

4142

4546

AGE OF ONSET Fig 3. The age of onset of arthritis, sex, and age distribution of the preceding. expanding skin lesion are shown. The median age of onset was I I . arthritis below age S was rare, I2 patients were adults, and the sex ratio was I .3: I in favor of males.

STEERE ET AL

12

Table 3. Prevalence Rates in Three Communities Number of Patients Town

Children Adults ~~

Total

Children

Adults

Total

3,942 1,246 3,518

5,400 1,600 4,900

9.6 22.6 12.3

0.5 6.4 0.5

3.0 10.0 3.9

8,706

11,900

12.2

1.4

4.3

14 8 17

2 8 2

16 16 19

1,458 354 1,382

Total

39

12

51

3, I94

a swimming place, or a particular food, could be identified. All but 3 families had either a dog or a cat.

Laboratory Findings None of the 51 patients was found t o be anemic or to have an elevated white blood cell count. Twenty patients with joint effusions at the time of the study (symptomatic patients) had a median erythrocyte sedimentation rate of 15 mm/hour (range: 4-55) compared to 8 mm/hour (range: 4-22) in 31 asymptomatic patients. Joint fluid cell counts on 9 patients showed a median white blood cell count of 26,000 cells/mm3 (range: 500-98,000), with a median of 87% polymorphonuclear leukocytes, 9% lymphocytes, and 4% tissue cells. The joint fluid had a median protein of 5 g% (range: 3.5-5.6), and the joint fluid glucose concentrations were not less than two-thirds those of the serum. In serum, none of the patients had a positive test for antinuclear antibodies, and only 1 had a positive test for rheumatoid factor (titer: 1 :20). In joint fluid, none of the 9 patients tested had antinuclear antibodies, but 4 had a positive test for rheumatoid factor (titer: 1 : 32 in all). The 20 symptomatic patients had a median C3 in serum of 78 mg% (range: 42-1 10 mg%) compared to 31 Table 4. Prevalence Rates on Four Roads ~

Number ofchildren Affected

Total

Old Lyme Road A Road B

4 4

51

East Haddam Road C Road D Total

Children Adults

Total

~

Old Lyme Lyme East Haddam

Town and Road

Prevalence per I ,OOO Residents

Population

Prevalence per 100 Children

asymptomatic patients who had a median C3 of 90 mg% (range: 70-100 mg%). (These distributions were not significantly different by the median test.) However 9 of 20 symptomatic patients had serum C3 levels below 70 mg% (the lower limits of normal in this laboratory), but none of 3 1 asymptomatic patients did (P< 0.005). Joint fluid C3 values on 9 patients were not less than twothirds of the serum C3 values. Only 2 of 20 patients (10%) were positive for the antigen HL-A B27, and both were adult women. Roentgenograms of affected joints in 18 patients showed only soft tissue changes. Five patients had synovial biopsies, and these specimens, stained with hematoxylin and eosin, showed synovial hypertrophy, vascular proliferation, and marked infiltration by lymphocytes, monocytes, and plasma cells. The results of serologic tests for various bacteria, mycoplasma, rickettsia, leptospira, and viruses are shown in Table 5. Although serum was taken from each patient only once, sometimes years after the onset of arthritis, and although a few patients had elevated titers against a particular organism, the serologies did not suggest that any of the agents tested caused arthritis in the 51 patients. Synovium and joint fluid from 5 patients were cultured for bacteria, mycoplasma, and viruses, and throat and rectal swab specimens from 7 patientsthose with the onset of arthritis during the study-were cultured for viruses. None of the specimens showed any evidence of infection.

DISCUSSION 65

7.8 7.7

5

26 33

15.4 12.1

17

I75

9.7

4

Thirty-nine children and 12 adults in three contiguous Connecticut communities were found to have an apparently similar type of arthritis. It was characterized by usually short but recurrent attacks of pain and swelling in a few large joints, often the knee, with longer intervening periods of complete remission and with, as yet, no permanent joint deformity. Most of the 39 chil-

LYME ARTHRITIS

13

Table 5. Antibody Titers* in 51 Patients ~~

~~

Agent Adenoviruses Arboviruses Group A Aura Ch ikungunya Eastern eq enceph Getah Highlands J Mayaro Mucambo Ndumu O'nyong-nyong Pixuna Ross River Semliki Forest Sindbis Una Western eq enceph Y-62-33 Group B Powasson St. Louis enceph Group other Bunyamwera Mermet Flanders Calif enceph

~~

Median (range) N t (N-64) N N N N N N N N N N N N N N N N N N N N N N

Agent Coxsackie viruses Group A 9 Group B I 2 3 4 5 Cytomegalovirus Hepatitis B (Ab$ and Ag) Herpes simplex Influenza Type A B Mumps Rubella Hemagglut inhibition IgM antibodies$ Leptospira (22 antigens)§ Gr A streptococci (ASLO) M pneumoniae Comp fixation Cold agglutination Rickettsia11 R akari R rickettsia R mooseri Coxiella burnetii

Median (range) N N N N 8 N N N N

(N-512) (N-8) (N-128) (N-32) (N-128) (N-32) (N-32) (N-64)

8 (4-32) 8 (4-32) 16 (N-32) 16 (N-256) N N 60 (60-120)

8 (N-256) N (N-32) N N N N

* Reciprocal t N-negative

of serum dilutions. undiluted. $ Tested in 16 patients instead of in all 51. 9 Tested in I I patients instead of in all 51. 1) Tested in all patients by complement fixation and in 10 patients by immunofluorescence.

dren studied had been thought initially to have juvenile rheumatoid arthritis. However many problems with that diagnosis emerged for both clinical and epidemiologic reasons. First, all but 1 patient had a monoarticular or an oligoarticular onset of their arthritis. In series from university hospitals, only 3 0 4 0 % of patients with JRA have had this type of onset, although such series are probably biased toward patients with more severe presentations (9-1 1). Second, oligoarticular JRA has an overall female-to-male ratio of 3 t o 1, and two peak ages of onset, 1-3 and 11-13 (9,ll-13). In this study there is no clear sex preponderance a t any age (Figure 3). Although the peak onset was at age l l , only l patient had the onset between ages 1 and 3, and 12 patients were adults. Third, 20-25% of children with the oligoarticular form of JRA develop antinuclear antibodies and/or iridocyclitis (10-13); none of the present patients has done so as yet. There is a subgroup of oligoarticular J R A consisting primarily of older children with negative tests

for antinuclear antibodies, many of whom eventually turn out to have ankylosing spondylitis (14). However, unlike the present patients, they are mostly males, sometimes develop acute iritis, and usually are HL-A B27 positive. The usual duration of joint pain and swelling in any one attack in the patients reported in this study was characteristically short, only 1 week, with much longer intervening periods of complete remission. This pattern has been described in JRA (15,16), but less frequently than one of exacerbations followed by partial remissions ( 16). The American Rheumatism Association defines JRA as a) 6 weeks of monoarticular or polyarticular arthritis if other symptoms such as intermittent fever, maculopapular rash, or morning stiffness are present, or b) 3 months of arthritis if they are not present (17). Only 23 of the patients reported here (48%) meet that definition. By the more rigorous definition of JRA of Ansell and Bywaters (18), which requires arthritis in 4 or more

STEERE ET AL

joints for 3 months in a child under age 16, only 15 of the patients reported here (31%) would qualify. But the major argument against the diagnosis of juvenile rheumatoid arthritis in these patients is their geographic, familial, and seasonal clustering. The prevalence of oligoarticular arthritis in the three communities is about 100 times the reported prevalence of JRA (19). Approximately 1 in 10 children living on four roads had the illness, and 6 families had more than 1 affected member. In addition, most patients had their onset in the summer or early fall. In summary, the diagnosis of juvenile rheumatoid arthritis in these patients is unlikely because of a) mono-oligoarticular onset of arthritis in almost all patients, b) the usual short duration ofjoint effusions, c) the lack of iridocyclitis or positive tests for antinuclear antibodies, d) the occurrence in 12 adults, e) the seasonal onset, and f) the prevalence of the disease in the three communities and within 6 families. Could some of these patients have had traumatic arthritis? Fifteen of them had only monoarticular arthritis for less than 3 months, and 2 of those had only one attack. However none gave a history of trauma, none appeared to have a structural abnormality of the affected joint, and some had associated symptoms such as fever, myalgia, or fatigue. All five synovial biopsies revealed marked infiltration with inflammatory cells, including three from patients with monoarticular involvement. The clustering of cases observed in this study suggests an infectious etiology. An outbreak of arthritis occurred in Haverhill, Massachusetts, in January 1926, which was transmitted by raw milk contaminated with Haverhillia multiformis (later called Streptobacillus moniliformis) (20,21). The onset of this illness was unusually abrupt, with severe chills, fever, vomiting, and headache, and the arthritis, which followed in many patients within days, was not recurrent and caused joint destruction in some. Thus the clinical characteristics of that illness are different from those in the patients reported here. Clustering of cases with Reiter’s syndrome has been observed following outbreaks of dysentery caused by Shigella (22,23). Arthritis following infection with this and other intestinal pathogens, Salmonella and Yersinia enterocolitica, has been linked to the histocompatibility antigen HL-A B27 (24-26). However none of these patients had diarrheal illnesses before the onset of arthritis, other clinical components of Reiter’s syndrome were absent, and only 2 of 20 tested were HL-A B27 positive; both were adult women. Although patients with rheumatic fever could conceivably cluster as observed

here, the arthritis in that entity is typically migratory, none of the patients met the Jones criteria for rheumatic fever (27), and none had an elevated antistreptolysin 0 (ASLO) titer. Arthritis may be associated with certain viral infections including rubella, rubeola, hepatitis B, mumps, smallpox, influenza, and some ECHO and adenoviruses (28.29). However arthritis is rarely the only or even the major symptom in these entities. The same is true for erythema infectiosum, an epidemic exantham of presumed viral etiology, which often causes arthritis in adults (30). The arthritis following vaccination for rubella usually affects large joints, often the knee, and short exacerbations and remissions may occur for years, much as observed here (31,32). However this type of arthritis follows 2 to 4 weeks after vaccination, and none of these patients was vaccinated within 1 year of the onset of arthritis. Three group A arboviruses-chikungunya, o’nyong-nyong, and Ross River-all transmitted through mosquito vectors, have been reported to cause primarily arthritis, in epidemic proportions (33-37). However chikungunya and o’nyong-nyong arthritides are dengue-like illnesses that typically affect large joints symmetrically, epidemic polyarthritis in Australia caused by the Ross River virus usually affects small joints in the hands, and none of the three has been reported to recur. Finally, the results of serologic tests and cultures in the present patients d o not suggest infection with rubella, group A arboviruses, or any other agent currently known to cause arthritis. The geographic clustering of the patients in more sparsely settled, heavily wooded areas rather than in town centers or along the shore, the peak occurrence in summer months, and the usual lack of close temporal onset in those living close together, are best explained by transmission of an agent by an arthropod vector or possibly by a continuing common source such as water. I n contrast, had the spread been from person to person or airborne, patients in close proximity such as those in the same family or the same classroom would have been expected to have had the onset of symptoms in the same year. The peak occurrence in the summer and early fall is compatible with an enterovirus infection originating possibly i n a contaminated lake or in contaminated drinking water. But the patients did not swim in a common place, all residents had their own wells, and throat and rectal swabs from those with the onset of arthritis during the study showed no evidence of viral infection. Thus the authors believe that the epidemiology fits best with an illness transmitted by an arthropod vector. I n this regard the occurrence of an unusual ex-

LY M E ARTHRITIS

panding skin lesion 1 to 24 weeks before the onset of arthritis in one-quarter of the patients is particularly intriguing. This lesion fits the description of an entity called erythema chronicum migrans, which was described by Lipschutz in 1913 (38). It has been reported in Europe, particularly in Scandinavia (38-44), but only recently in the United States, interestingly in eastern Connecticut (45,46). The disease is thought to be transmitted by the sheep tick, Ixodes ricinus (39), but this vector has been disputed because the disease has been reported in areas that do not have ticks (44). Although allergic and toxic reactions to the bite have been postulated as a cause of the lesion, the most widely held view is that an infectious agent is involved. Binder et al were able to reproduce the disease serially in volunteers by inoculation of material from the edge of the lesion (47); these results were reproduced by Sonck, who induced a ring on his own forearm (43). In 1962 two French groups reported that some patients with the illness have positive microagglutination titers against Rickettsia conori, R mooseri, and R prowazeki (42), but others have not been able to confirm this finding. The lesion, which may not become apparent for weeks or months after the bite, may last from a few days to many months (38-44). Associated symptoms include fever, malaise, fatigue, and headache. Several authors have noted that neurologic symptoms such as meningeal and encephalitic signs, hyperesthesias, and peripheral neuropathy may occur (40), but others think that these symptoms may have resulted from other diseases that occurred coincidentally (48). Arthritis, however, has not been associated with erythema chronicum migrans, and-if this is indeed the lesion that the present 13 patients experienced-there can be no certainty that it is related. However only 2 of 159 family members of these 51 patients had the skin lesion without developing arthritis. Although the prevalence of erythema chronicum migrans in the three communities reported here is not known, it is doubtful that it is so common that one-quarter of the patients with arthritis would have the skin lesion just weeks before the arthritis by chance alone. Thus the authors think that both symptoms may be manifestations of the same illness. The best treatment for this illness is not clear. Some physicians have reported that penicillin or tetracycline results in disappearance of the skin lesion (41,42), but others find antibiotics ineffective. Four of the patients with expanding skin lesions received penicillin but still developed arthritis. Because most of the patients with arthritis were thought to have JRA, most were

15

treated with aspirin. But regardless of aspirin, other antiinflammatory agents, immobilization, aspiration of the effusion, or no therapy, joint attacks were usually short. However 2 patients with persistent effusions underwent aspiration and injection of corticosteroid esters, only to have the effusions return within days; after six months 1 of them had a synovectomy. Salicylates do not seem either to suppress the more prolonged attacks or to prevent recurrences. At present, we usually prescribe aspirin only during symptomatic periods. The authors believe that the arthritis described here is a previously unrecognized clinical entity and have named it “Lyme arthritis,” after the community where it was first studied. Can it be distinguished from oligoarticular juvenile rheumatoid arthritis-another disease(s) of unknown etiology-in the individual patient? Perhaps, if the characteristic skin lesion has been present. Otherwise the usual short duration of joint effusions, the typically complete remissions, the lack of antinuclear antibodies .and iridocyclitis, and possibly low serum C3 levels during symptomatic periods are somewhat suggestive, but certainly not diagnostic, of Lyme arthritis. Because other patients from eastern Connecticut who fit the clinical description have been seen, it is believed that Lyme arthritis extends beyond the three communities studied here; but how far beyond is not known.

ADDENDUM I n the 1976 summer season-after this article was completed-the authors saw 38 new patients with Lyme arthritis. In this group we have a) confirmed prospectively the importance of erythema chronicum migrans as a marker for subsequent Lyme arthritis, and b) found serum cryoprecipitates associated with clinical activity of skin and joints (49,50).

ACKNOWLEDGMENTS The authors thank the numerous primary and consulting physicians who supplied information about their patients, and local health and school officials in Old Lyme and East Haddam who supplied census data and follow-up support for the study. We gratefully acknowledge the laboratory assistance of Dr. Dorothy M. Horstmann and Ms. Jean Emmons for rubella serologies, Ms. Catherine R. Sulzer for leptospira serologies, Dr. George L. LeBouvier for hepatitis testing, Dr. Edward S. Murray for rickettsia1 testing by immunofluorescence. Ms. Grace Tucker for viral cultures, Ms. Johanna Shansky for tissue typing, Dr. Alexander Baumgarten for

STEERE ET AL

16

rheumatic disease serologies, Dr. John Redys for ASLO titers, and Dr. Philip W. Askenase for microscopy reports. We also thank Mr. Daniel Freeman for statistical consultation, Mrs. Stella Cretella for technical assistance, and Mrs. Wanda D. Prinz for preparation of the manuscript.

REFERENCES I. Steere AC, Malawista SE, Snydman DR, et al: A cluster of arthritis in children and adults in Lyme, Connecticut. Arthritis Rheum 19:824, 1976 (abstr) 2. Casals J: Immunological techniques for animal viruses, Methods in Virology. Edited by K Maramorosch, H KOprowski. New York, Academic Press, 1967, pp 113-198 3. Clarke DH, Casals J: Techniques for hemagglutination and hemagglutination-inhibition with arthropod-borne viruses. Am J Trop Med Hyg 7561-573, 1958 4. Sulzer CR, Jones WL: Leptospirosis: Methods of Laboratory Diagnosis. HEW Publication No. C D C 76-8275 5 . Liebhaber H: Measurement of rubella antibody by hemagglutination inhibition. J lmmunol 104:826-834, 1970 6. Ankerst J, Christensen P, Kjllen L, et al. A routine diagnostic test for IgA and IgM antibodies to rubella virus: absorption of IgG with Staphylococcus aureus. J Infect Dis 130:268-273, 1974 7. Melnick JL, Wenner HA: Diagnostic Procedures for Viral and Rickettsia1 Infections. Fourth edition. New Y ork, American Public Health Association, 1969, p 589 8. Terasaki P, McClelland J, Park MS, et al: Microdroplet lymphocyte cytotoxicity test, Manual of Tissue Typing Techniques. Edited by J Ray, D Hare, D Kayhoe. US Government Printing Office, DHEW (NIH 74-545), July, 1973, p 54 9. Calabro JJ, Holgerson WB, Sonpal G M , et al: Juvenile rheumatoid arthritis: a general review and report of I 0 0 patients observed for 15 years. Semin Arthritis Rheum 5:257-298, 1976 10. Schaller J, Wedgwood RJ: Juvenile rheumatoid arthritis: a review. Pediatrics 50:940-953, 1972 1 I . Rudnicki R D , Ruderman M, Scull E, et al: Clinical features and serologic abnormalities in juvenile rheumatoid arthritis. Arthritis Rheum 17:1007-1015, 1974 12. Petty RE, Storm PB, Cassidy JT, et al: Immunologic correlates of antinuclear antibodies in juvenile rheumatoid arthritis. Arthritis Rheum 12:323-324, 1969 13. Calabro JJ, Parrino GR, Atchoo PD, et al: Chronic iridocyclitis in juvenile rheumatoid arthritis. Arthritis Rheum 13:406-413, 1970 14. Schaller JG, Ocho H D , Thomas ED, et al: Histocompatibility antigens in childhdod-onset arthritis. J Pediatrics 88:926-930, 1976 15. Calabro JJ: The early natural history of juvenile rheumatoid arthritis: a 10-year follow-up of 100 cases. Med Clin N Am 52567-591, 1968 16. Brewer El Jr: Juvenile rheumatoid arthritis, Major Prob-

lems in Clinical Pediatrics. Vol. 6. Philadelphia, Saunders, 1970, p 12 17. biagnostic and Therapeutic Criteria Committee of the American Rheumatism Association: Criteria for the classification of juvenile rheumatoid arthritis. Bull Rheum Dis 23:712-719, 1973 18. Ansell BM, Bywaters EGL: Prognosis in Still’s disease. Bull Rheum Dis 9:189-192, 1959 19. Cobb S: The prevalence and incidence ofjuvenile rheumatoid arthritis, The Frequency of Rheumatic Diseases. Cambridge, Mass, Harvard University Press, 1971, pp I 1 1 - 1 12 20. Place EH, Sutton LE: Erythema arthriticum epidemicum (Haverhill fever). Arch Intern Med 54:659-684, 1934 21. Farrell E, Lordi G H , Vogel J: Haverhill fever. Arch Intern Med 64:l-14, 1939 22. Noer HR: An “experimental” epidemic of Reiter’s syndrome. JAMA 198:693-698, 1966 23. Paronen I: Reiter’s disease. Acta Med Scand [Suppl] 212:7-112, 1948 24. Ahvonen P, Sievers K , Aho K: Arthritis associated with Yersinia enrerocolitica infection. Acta Rheumatol Scand 15:232-253, 1969 25. Aho K, Ahvonen P, Alkio P, et al: HL-A 27 in reactive arthritis following infection. Ann Rheum Dis 34:29-30, 1975 (suppl) 26. Calin A, Fries JF: An “experimental” epidemic of Reiter’s syndrome revisited. Ann Intern Med 84564-566, 1976 27. Stollerman G H , Markowitz M, Taranta A, et al: Jones’ criteria (revised) for guidance in the diagnosis of rheumatic fever. Circulation 32:664-668, 1965 28. Panush RS: Adenovirus arthritis. Arthritis Rheum 17534-536, I974 29. Smith JW, Sanford JP: Viral arthritis. Ann Intern Med 67~651-658, 1967 30. Ager EA, Chin TDY, Poland JF: Epidemic erythema infectiosum. N Engl J Med 275:1326-1331, 1966 31. Thompson G R , Ferreyra A, Brackett RG: Acute arthritis complicating rubella vaccination. Arthritis Rheum 14: 19-26, 1971 32. Thompson G R , Weiss JJ, Shillis JL, et al: Intermittent arthritis following rubella vaccination. Am J Dis Child 1251526-530, 1973 33. Lumsden WHR: An epidemic of virus disease in Southern Province, Tanganyika Territory, in 1952-53. 11. General description of epidemiology. Trans R SOCTrop Med Hyg 49:33-57, 1955 34. Ross RW: The Newala epidemic. I I I . The virus: isolation, pathogenic properties and relationship to the epidemic. J Hyg (Camb) 54:177-191, 1956 35. Haddow AJ, Davies CW, Walker AJ: O’nyong-nyong fever: an epidemic virus disease in East Africa. I. Introduction. Trans R SOCTrop Med Hyg 54517-522. 1960 36. Williams MC, Woodall JP, Gillett JD: O’nyong-nyong fever: an epidemic virus disease in East Africa. VII. Virus isolation from man and serological stbtiies up t o July

LYME ARTHRITIS

1961. Trans R SOCTrop Med Hyg 59:186-197, 1965 37. Seglenieko Z, Moore BW: Epidemic polyarthritis in South Australia. Med J Aust 2:552-556, 1974 38. Lipschutz B: Quoted in Dermatology in General Medicine. Edited by TB Fitzpatrick et al. New York, McGrawHill, 1971, pp 709-71 I 39. Afzelius A: Erythema chronicum migrans. Acta Derm Venereol (Stockh) 2:120, 1921 40. Hellerstrom S: Erythema chronicum migrans. Acta Derm Venereol (Stockh) I1:315, 1930 41. Flanagan BP: Erythema chronicum migrans Afzelius in Americans. Arch Dermatol 86:4 10-41 1, 1962 42. Degos R, Touraine R, Arouete J: L’erythema chronicum migrans. Ann Dermatol Syphiligr (Paris) 89:247, 1962 43. Sonck CE: Erythema chronicum migrans with multiple lesions. Acta Derm Venereol (Stockh) 4534-36, 1965 44. Hard S: Erythema chronicum migrans associated with a mosquito bite. Acta Derm Venereol (Stockh) 46473-476, 1966

17

45. Scrimenti RJ: Erythema chronicum migrans. Arch Dermatol 102:104-105, 1970 46. Mast WE, Burroughs W: Erythema chronicum migrans in the United States. JAMA 236:859-860, 1976 47. Binder E, Doepfmer R, Hornstein 0: Quoted in Dermatology in General Medicine. Edited by TB Fitzpatrick et al. New York, McGraw-Hill, 1971, pp 709-711 48. Putkonen T, Mustakallio K K , Salminen P: Erythema chronicum migrans with meningitis: a rare coincidence of two tick-born diseases. Dermatologica 125:184-1 88, 1962 49. Steere AC, Hardin JA, Malawista SE: Lyme arthritis: serum cryoprecipitates associated with skin and joint lesions. Arthritis Rheum 20: 136, 1977 50. Steere AC, Hardin JA, Malawista SE: Erythema chronicum migrans and Lyme arthritis: serum cryoprecipitates associated with clinical activity of skin and joints. Submitted for publication